Many wireless communication systems transmit special signals, sometimes referred to as training signals, to allow a receiver to acquire the signal, to perform automating gain control (AGC), to perform carrier frequency and timing corrections, and to estimate the characteristics of the radio-frequency (RF) channel. A multiple-input multiple-output (MIMO) communication system may use two or more antennas to transmit data and two or more antennas to receive data. Channel estimates in these MIMO systems may be used for channel equalization and to separate different spatial streams transmitted together. In some cases, channel estimates may also be used for beamforming. In some MIMO systems, such as MIMO systems in accordance with the World Wide Spectrum Efficiency (WWiSE) proposal for a high-throughput extension to the Institute of Electrical and Electronics Engineers standards for wireless local area networks discussed in more detail below, interpolation may be required to determine the channel characteristics for all subcarrier frequencies in the frequency bandwidth because of the configuration of the training signals. This interpolation generally assumes that the channel characteristics are smooth, at least with respect to adjacent or nearby subcarriers. In many cases, the channel is not smooth resulting in less accurate channel estimates. This makes channel equalization and the separation of spatial streams in MIMO systems difficult. It may also result in the generation of less-accurate beamforming coefficients.
Thus there are general needs for methods for generating channel estimates for a MIMO channel that do not require interpolation.